Version 1.16 candidate 1 released

We are happy to announce the availability of OpenTURNS 1.16 RC1 today.

This release introduces grid graphs, hdf5 storage, a new example gallery:

… and more under the hood.

For detailed informations you can have a look at the changelog:

Or the release highlights:

For debian users, note the debian repository moved ant that you may have to update your configuration, see the installation instructions.

This a release candidate, your feedback is important for the final release due in a few weeks,
so dont hesitate to report bugs.

The OpenTURNS team


I just read the changelog and discover the FittingTest::Lilliefors that I did not know. I read that it is more robust that Kolmogorov when parameters’ distribution are estimated from the sample. But it seems that it is only valid when the sample is from a Normal population, is this correct ?

Hey @schueller thanks for the release highlights notebook!
I had it in mind and just started this but this is more complete :ok_hand:

@dumas The Lilliefors is valid for all factories. Have a look at Lilliefors Test

1 Like


I have the openturns 1.15 installed in my system, how can I get (or updated) the latest version of Openturns via anaconda, using a command line?

Thanks in advance,

Hello @oussama-braydi. Normally the install command will update your package if it is already installed. Version 1.16 is the current stable version of OpenTURNS.

conda install -c conda-forge openturns

1 Like

Thank you @josephmure for your answer,
However, when I try to update or install openturns in the current conda environment, many conflicts are displayed.
The problem is fixed by creating a new conda environment and use it to install openturns 1.16 and the other needed package.

1 Like

Hi Antoine!
The Kolmogorov test is not more robust when parameters are estimated: Kolmogorov test is wrong when parameters are estimated from data.
The question has already been discussed at:


This might be interesting to know for Persalys :slight_smile:
Best regards,